Abstract :
Background: AI-assisted mental health interventions are becoming more prevalent in Pakistan; however, it is not understood how the technology can foster safe (calibrated) trust versus unhealthy dependence among young people, or what psychological factors drive the use of AI interventions so that they are used alongside human help. Objective: Psychological predictors of safe trust, unhealthy dependence, and active help-seeking towards AI mental health support among young Pakistanis. Methods: Data collection methods involved a cross-sectional survey of 402 young people (aged between 18 and 30 years) of which 58.2% were female. Various scales including trait anxiety, interpersonal trust propensity, perceived stigma, digital self-efficacy, help-seeking experiences, and outcomes were given to participants. Data were analyzed using regression, moderation, and mediation analyses. Results: Digital self-efficacy (? = .42), trust propensity (? = .21), and stigma – (? = –.17) were found to be significant predictors of safe trust. Trait anxiety (? = .48), prior negative help-seeking (? = .26), and low digital self-efficacy (? = –.15) were all predictors of unhealthy dependence. Safe trust (? = .41), reduced stigma (? = –.24), and negative prior experiences (? = –.14) were all significant predictors of active help-seeking. Anxiety is a more significant predictor of dependence than of trust or help-seeking (p < .001). The trust–help-seeking relationship was moderated by collectivism (? = .14, p = .003) and self-efficacy–trust was moderated by digital access inequality (? = –.18, p = .001). There was no sign of any decrease in religiosity. Conclusion: Digital self-efficacy, low stigma and trust propensity are factors that help young Pakistanis develop safe trust. Negative experiences and anxiety heighten dependence risk. Safe trust encourages integrated help-seeking – especially in collectivist families. Interventions to foster digital literacy, decrease stigma, monitor users who are experiencing digital anxieties, and address digital access inequality.